Materials Today Communications 22 (2020) 100755

Contents lists available at ScienceDirect

Materials Today Communications

journal homepage: www.elsevier.com/locate/mtcomm

Use of ball mill to prepare nanocellulose from eucalyptus biomass: T

Challenges and process optimization by combined method
Rafaela R. Ferreiraa, Alana G. Souzaa, Lucas L. Nunesa, Naresh Shahib, Vijaya K. Rangaric,
Derval dos Santos Rosaa,*
a
Centro de Engenharia, Modelagem e Ciências Sociais Aplicadas – CECS/ Universidade Federal do ABC (UFABC), Avenida dos Estados, 5001, CEP: 09210-580, Santo
André, SP, Brazil
b
Integrative Biosciences (IBS) Graduate School of PHD Program, Tuskegee University, AL 26088, USA
c
Department of Materials Science and Engineering, Tuskegee University, Tuskegee, AL 36088, USA

A R T I C LE I N FO A B S T R A C T

Keywords: A combination of mechanical methods was developed to obtain a suspension of nanocellulose (NC) from
Mechanical methods Eucalyptus sawdust. In this two-step process, the ball milling (1–4 h) was followed by a high-intensity ultrasound
Cellulose nanostructures (10–30 min) irradiation. The prepared samples were characterized using Fourier-transform infrared spectro-
Ball mill scopy, Raman, dynamic and static light scattering, Zeta potential, X-ray diffraction, X-ray photoelectron spec-
High-intensity ultrasound
troscopy, Thermogravimetric analysis, Field-emission scanning electron microscopy, and Transmission electron
Eucalyptus sawdust
microscopy. The crystallinity was calculated using different methods and compared. The mechanical grinding
Nanoparticles’ characterization
results in a random and irregular break of the fibers. In a combined method, the ultrasound waves generate more
defibrillation and separation of the nanofibers and, consequently, the reduction of fiber diameters. The high-
intensity ultrasound promoted an increase in the homogeneity, crystallinity, and electrostatic and thermal sta-
bility of the NC suspension. The samples milled for 2 h and 20 min ultrasound irradiation presented the high-
efficiency in obtaining NC.

1. Introduction
Cellulose is an environmentally friendly biopolymer derived from
natural sources such as forest waste, woody biomass, crops, algae, and
microorganisms. Cellulose and its derivatives are sustainable and bio-
degradable, therefore receiving exceptional attention as a substitute for
petrochemical feedstock in polymer industries [1]. More recently, uses
of cellulose from the forest residues have been devoted to the isolation
of nanocellulose (NC), and Eucalyptus sawdust is one of them. Eu-
calyptus is a fast-growing wood that covers a larger planted area with
approximately 187 million ha in the world [2]. During the processing of
eucalyptus, an enormous amount of sawdust is originated, which is rich
in lignocellulosic constituents ˜25−35 wt.% of lignin, ˜20−30 wt.% of
hemicellulose and ˜35−40 wt.% of cellulose [3]. The residues are often
disposing as waste or burning for energy production. It results in high
emissions of pollutant gas [4] leads to undesirable environmental
consequences.
Nanocellulose is a new class of eco-material with unique properties
such as nanoscale dimension, high surface area, high strength, specific
modulus, unique morphology, and excellent optical properties [5]. Due
to the unique properties of NC, this material is applicable in many
fields, such as mechanical reinforcements in composite materials,
bioengineering, pharmaceutical, and biomedical technologies [6]. A
wide range of applications and products can be developed; therefore,
NC is a fascinating next-generation engineering material to researchers
and industries.
Several top-down approaches are available for NC isolation, such as
chemical, biological, mechanical, and combination of chemical and
mechanical [7–9]. However, an environmentally-friendly combination
method that does not employ strong acids or toxic reagents has not been
widely explored. Various chemical processes are required to transform
the lignocellulose rich sawdust into cellulose pulp [10].
Usually, chemical methods are conducted by acid hydrolysis using
strong acids, and it is well established [5,11–14]. However, chemicals
are always risky to handle, corrosive to the equipment’s, and detri-
mental effect on the environment [5,10]. The enzymatic hydrolysis
represents the biological method. This method requires long reaction
times, and the cost of enzymes is usually high and also limited uses

⁎
Corresponding author.
E-mail addresses: reis.f@ufabc.edu.br (R.R. Ferreira), alana.gabrieli@ufabc.edu.br (A.G. Souza), lucas.nunes11@fatec.sp.gov.br (L.L. Nunes),
nshahi1164@tuskegee.edu (N. Shahi), vrangari@tuskegee.edu (V.K. Rangari), dervalrosa@yahoo.com.br (D.d.S. Rosa).

https://doi.org/10.1016/j.mtcomm.2019.100755
Received 3 August 2019; Received in revised form 28 October 2019; Accepted 5 November 2019
Available online 07 November 2019
2352-4928/ © 2019 Elsevier Ltd. All rights reserved.
R.R. Ferreira, et al.
[15,16]. On the other hand, mechanical methods are simple and do not
require chemical reagents; this type of method are outstanding due to
their high production capacity and low cost [17–19]. The main ex-
amples of mechanical methods are ball mill, cryogenic break, and ul-
trasound.
The ball mill is an environmentally friendly and low-cost technique
that is widely applied in the industry all over the world. It consists of a
cylindrical shell rotating around its axis, which is partially filled with
balls; the energy released from impact and attriction between the balls
and the samples converts the particles from micro- to nanoscale [20].
However, some disadvantages include the possibility of contamination,
usually ceramic or metal, from grinding medium and container, the
formation of irregular shapes, and extended milling and cleaning times
[5,21]. The current literature survey shows the frequent uses of pla-
netary ball milling (high-energy ball mill) or shaker mill for obtaining
NC, and/or acid solutions during the milling. The planetary ball mill
achieves higher rotations and requires much energy; it has very low
milling capacity due to the small containers and is not common on an
industrial scale [5,17,22]. The tumbler ball mill, which was used in this
work, is a very simple process that depends mainly on the time and
balls and diameter of the container [21,23,24].
While the ball mill involves friction and shear forces, the ultrasound
process applies the use of oscillating power to create hydrodynamic
forces of ultrasound [25]. The high intensive waves result in the for-
mation, expansion, and implosion of microscopic gas bubbles when the
sample absorbs the energy, and this intense energy decreases the size of
the particles [26,27]. However, generally high times or low sample
quantity per process need to be used. Aiming to find an environmental
process with high efficiency and low cost, researches are combining two
or more isolation methods at the same time or in sequence. The com-
bination of mechanical methods, such as ball mill and high-intensity
ultrasound have recently been highlighted [22].
The purpose of the study was to develop a combined mechanical
method (ball milled and ultrasound) to isolated NC from Eucalyptus
sawdust. This method can be extended to large scale production, which
is different from the existing literature, is simple, low cost, and com-
mercially attractive. It is important to mention that the production of
nanocellulose in a pilot-scale is currently carried out mainly by acid
hydrolysis [28,29]. The NCs isolated were characterized by morpho-
logical, chemical, physical, and thermal properties. Furthermore, the
crystallinity of the NC was compared with XRD, Raman, and FTIR.
Therefore, this study aimed to transform Eucalyptus residues as pro-
mising sources of high value-added biopolymer (in the NC form).
2. Experimental
2.1. Materials
Eucalyptus citriodora waste (eucalyptus sawdust – ES) was kindly
donated after harvesting and cutting at the state of Mato Grosso, Brazil.
The composition of the eucalyptus sawdust was ˜33 wt.% of lignin,
˜27 wt.% of hemicellulose, and ˜40 wt.% of cellulose. All the reagents
were purchased from Sigma-Aldrich (SP-Brazil) (NaClO2, 99 %, NaOH
-95 %, KOH-98 %, ethanol–98 %).
2.2. Methods
2.2.1. Treatments
To obtain cellulose without the presence of lignin and hemi-
cellulose, two chemical pre-treatments were carried out in sequence:
2.2.1.1. First treatment. The ES was first treated in a solution of NaClO2
(3.8 wt.%) for 2 h at 50 °C under mechanical stirring, and then the
obtained material was washed and dried in an oven at 50 °C for 12 h.
2.2.1.2. Second treatment. Subsequently, the material was treated in a
2
Materials Today Communications 22 (2020) 100755
solution of NaOH (10 wt.%) and KOH (10 wt.%) for 2 h under magnetic
stirring, washed and dried at 50 °C for 12 h. After these treatments, the
final material has in its composition ˜1.5 wt.% of hemicellulose and
˜2 wt.% of lignin, and this sample is named T2. This methodology was
previously reported [30]. The yield was ˜40 wt.%. Considering that the
eucalyptus residues have ˜40 wt.% of cellulosic fibers and ˜60 wt.% of
other non-cellulosic components, this yield is considered as excellent.
2.2.2. Isolation of nanocellulose
The treated sample, after the second treatment, was mixed with
ethanol (80 % v/v) in the proportion 1:1 g/ml and subjected to 4 dif-
ferent milling times: 1 h, 2 h, 3 h, and 4 h, in a ball mill (MA500 -
Marconi). This is rotational type mill, and the particles are shrunk by
frictional and friction forces that occur between the balls, the jar wall,
and the sample during the rotational movement of the jar. The speed of
the milling is 200 rpm. The jars used are made of porcelain, model 21A
(Chiarotti), with the following characteristics: diameter of 150 mm,
height of 195 mm, capacity of 1 L. The jar had 70 % of its volume filled
with balls and 30 % with samples. The jar was filled with 5 g of treated
sample, 5 ml of ethanol, 305 g of larger alumina balls (diameter of
21 mm), and 44 g of smaller alumina balls (12 mm). The balls are 92 %
pure, the other 8 % being silica or silicon dioxide. The ball mill con-
ditions were selected according to our previous studies [20,31]. A
photograph of the jar and balls used is presented in Fig. 1 S (Supple-
mentary material). The production yield of the ball-milled product is
˜80 wt.% (this value was calculated comparing the sample added in the
jar of ball mill and the sample obtained after the milling time).
After 1 h interval (each hour) of milling, the samples were oven-
dried at 35 °C for 24 h to remove the ethanol. Then the samples in
powder form were subjected to high-intensity ultrasonic treatment in a
Sonics Vibra Cell, at the power of 400 W and 24 kHz, in a continuous
process. During the ultrasonication, the samples were maintained under
an ice bath. The tip diameter is 14 mm, with an amplitude of 20 %. For
each sample, aqueous solutions containing 1 wt.% of NC were sonicated
during the selected time: 10, 20, and 30 min. After the ultrasound, the
samples were in suspension form (1 wt.%). The image presented in
Fig. 1 represents the followed methodology in this study, and Table 1
shows the adopted nomenclatures.
2.2.3. Characterization
2.2.3.1. Fourier transform infrared spectroscopy. Fourier transform
infrared spectroscopy (FT-IR) was conducted using a Frontier 94,942
(PerkinElmer, USA) in the range of 4000 to 400 cm-1 with 32 scans and
a resolution of 4 cm-1 with attenuated total reflectance (ATR) accessory
equipped with diamond. The FTIR relative crystallinity (index of
crystallinity - IC) was calculated using the ratio of the absorption
bands at 1369 and 1315 cm-1, as reported by Lee et al. (2015) [32].
Pellet pressed samples were analyzed at room temperature.
2.2.3.2. Raman spectroscopy. All samples were analyzed with a Raman
spectroscopy (XRD Raman Spectroscopy Thermo Scientific, Madison,
USA). This system is equipped with a 780 nm wavelength with 5.0 mW
power laser. Approximately 0.5 mm–1 mm length of each dried sample
is analyzed at 50x magnification. The spectra were recorded, and each
spectrum presented is an average of 3 different points on the sample.
The data acquisition was carried out in a range of 100-3500 cm−1, and
32 scans were accumulated in each sample with 4 cm-1 spectral
resolution. All the graphs were baseline corrected (25 points) using
OriginPro 2009. From the subtracted spectrum of the NCs, peak
intensities for the 380 and 1096 cm-1 bands were measured. The
intensity ratio I380/I1096 was determined, and Raman crystallinity
(index of crystallinity – IC) was calculated using Eq. 1 [32–34].
[(I380 − I1096)) − 0.0286]
IC =
0.0065 (1)
R.R. Ferreira, et al. Materials Today Communications 22 (2020) 100755

Fig. 1. Flow diagram representative of the preparation of nanocellulose: the pretreatment methods, followed by the mechanical grinding (ball mill) and finishing
with assisted ultrasound methodology.

2.2.3.3. Dynamic and static light scattering. The isolated NC
measurements were performed using dynamic light scattering (DLS)
with a stable 90° scattering angle. Static light scattering (SLS) was
conducted in the region 30–150 °. Both samples were performed using
the ALV/CGS-3 instrument consisting of a polarized HeNe laser
(22 mW). The equipment was operating at a wavelength (ʎ) 633 nm,
with NC diluted to a concentration of 0.01 wt%. The reading time
established to obtain the DLS data was 10 s, and the measurements were
performed in triplicate using the final average values. Hydrodynamic
radius (RH) was calculated by using Eq. 2 [35]:
KB T
RH =
3πηD
Where: KB is the Boltzmann's constant, T is the temperature, η is the
viscosity, and D is the translational diffusion coefficient.
In the SLS experiment, the angular dependence of the absolute ex-
cess time-averaging (for a duration of scanning of 10 s) scattering in-
tensity, known as the Rayleigh ratio Rvv (θ) leads to the mean square
radius of gyration (RG), which allows identification of the geometric
shape of the objects (RG/RH) [36].
2.2.3.3.1. Morphological analysis – field emission scanning electron
microscopy and transmission electron microscopy (FE-SEM and TEM). The
surface morphology of all the samples was analyzed at different stages
of treatments using FE-SEM (JEOL-JSM-7200 F) microscope equipped
with energy-dispersive X-ray spectroscopy (EDS), with an acceleration
voltage of 10 kV. The sample was prepared; drop of the NCs (2 wt.%)
was dropped in the stub and after dried in the stove at 60 °C during 2 h.
or Samples were mounted on carbon conductive tape and sputtered for
around 5 min in Hummer 6.2 sputter system under 10 mA.
The microstructure and morphology of the samples with the best
conditions were investigated by Transmission Electron Microscope
(TEM-Joel 2010). A drop of the colloidal solution at 2 wt.% was de-
(2)
posited on a copper grid for analysis.
2.2.3.4. Zeta potential. The zeta potential measurement was
determined using Zetasizer Nano-ZS (Malvern Instruments) to
evaluate electrophoretic mobility of particles converted to Zeta
potential using the Smoluchowski [30] Eq. 3. The reading time
established to obtain the zeta potential data was 10 s, and the
measurements were performed in triplicate using the final average

Table 1
Nomenclatures of samples prepared by varying milling and high-intensity ultrasonic times.
SAMPLE NOMENCLATURE SAMPLE NOMENCLATURE

1M 1h milling 3M 3h milling
1 M-10U 1h milling - 10 min ultrasound 3 M-10U 3h milling - 10 min ultrasound
1 M-20U 1h milling - 20 min ultrasound 3 M-20U 3h milling - 20 min ultrasound
1 M-30U 1h milling - 30 min ultrasound 3 M-30U 3h milling - 30 min ultrasound
2M 2h milling 4M 4h milling
2 M-10U 2h milling - 10 min ultrasound 4 M-10U 4h milling - 10 min ultrasound
2 M-20U 2h milling - 20 min ultrasound 4 M-20U 4h milling - 20 min ultrasound
2 M-30U 2h milling - 30 min ultrasound 4 M-30U 4h milling - 30 min ultrasound

3
R.R. Ferreira, et al.
values. The NCs were diluted to a concentration of 0.01 % wt.%.
3 η UE
ζ=
2 ε f (ka) (3)
2.2.3.5. X-ray diffraction. X-ray powder diffraction data were collected
on a D8 Focus diffractometer (Bruker AXS - Karlsruhe, Germany),
operating at 40 kV and 40 mA, with monochromatic CuKα1 radiation
(λ =1.54056 Å) selected by a curved Ge (111) crystal monochromator.
The data were collected from 10° to 60° in steps of 0.01° and a counting
time of 100 s at each 0.5°. The index of crystallinity (IC) was calculated
according to the following equation (Eq. (4)):
IC(%)= (I200-Iam)/I200 *100 (4)
Where I200 is referred to the maximum intensity of the 200 lattice dif-
fraction peak at 2θ = 22.2° and Iam is the intensity of diffraction for the
amorphous part at 2θ = 18.5° [37,38].
2.2.3.6. X-ray photoelectron spectroscopy. The XPS analysis was
conducted on a ThermoFisher Scientific, model K-alpha+, equipped
with a monochromatic A1 Kα radiation at 1486.6 eV. The analysis was
performed at room temperature and operated at 10 or 20 eV for high-
resolution scans. Data were treated with CASAXPS software.
2.2.3.7. Thermogravimetric analysis. Thermogravimetric analysis (TGA)
was carried out using STA 6000 instrument (PerkinElmer, USA). The
samples were heated from 20 to 600 °C with a constant heating rate of
20 °C/min under nitrogen flow. All the graphs were analyzed using the
software OriginPro 2009.
3. Results and discussion
3.1. Chemical characterization of the eucalyptus biomass and its
nanocelluloses
A basic characterization of the chemical structure of the samples
with different grinding and ultrasound times was conducted for a pre-
vious evaluation of the fiber composition and apparent crystallinity
using FTIR and Raman spectroscopy. Fig. 2a presents the full spectra of
the milled samples.
All the samples showed similar spectra, which is indicative of the
similarity between the chemical composition of the samples [22]. It was
observed bands of 3500-3200 and 2900 cm-1, characteristic of the −OH
binding and C–H symmetrical stretching groups of cellulose [31,39].
The bands around 898, 1030, 1160, 1315, 1369, and 1425 cm-1 were
assigned to the C–H vibrations, C–O stretching, C–O-C asymmetric vi-
bration, −CH2 rocking vibration and −CH2 deformation vibration of
the cellulose, respectively [40].
Fig. 2a also shows a zoom of the region between 1800 and 1400 cm-
1
; it is evident that all the milled samples do not show the peaks related
to the lignin and hemicellulose [40,41]. Other peaks assigned to the
cellulose structure are found at 1155, 1060, 1010, 990, and 893 cm-1.
Most of these peaks are influenced by neighboring bonds and their vi-
brational modes, which results in high band overlap and shift of some
peaks, especially below 1100 cm-1. After the milling, it was noted a
significant decrease in the intensity of the band 1640 and
3400−3000 cm-1 due to the increase in the apparent crystallinity of the
samples; increasing chain ordering, there is less available spacing for
water entering the sample. Functional groups between 1300 to 800 cm-1
were observed; the increasing intensity of these groups indicates in-
creased crystallinity of cellulose in the milled samples.
The apparent index of crystallinity (IC) was calculated for the
samples 1 M, 2 M, 3 M, and 4 M. The sample ES showed an overlap of
peaks due to the presence of non-cellulosic compounds, and it was not
possible to calculate the "apparent" IC, as presented in Table 2 [16,32].
4
Materials Today Communications 22 (2020) 100755
This value is not the real value and was used only for comparing the
samples with each other [32].
The ball mill is a challenge composed of three main aspects: (i)
energy and time, (ii) size and stability of nanoparticles, and (iii) wear:
contamination and amorphization (Fig. 3).
In general, process optimization needs to balance the three chal-
lenges for shorter times and energy, smaller particle sizes, high stabi-
lity, and crystallinity. According to the crystallinity values, it is not
possible to unite all the desired characteristics in one material only with
the ball mill: the sample 2 M showed high crystallinity (however, as
presented in the next sections, the sizes were larger than the desire for
this type of nanoparticles).
Less crystallinity of the cellulose obtained from samples 1 M, 3 M
and 4 M are justified by the inefficient break resulting in the impact and
shear forces or to the excessive time that resulted in the breaking of
both the amorphous and crystalline regions. Combine mechanical
method is an alternative to use low times in the ball mill and guarantee
the aimed properties in the dimensional and stability characteristics of
the nanocellulose.
The comparison between the FTIR spectra of the milled samples and
its ultrasonicated samples is available in Fig. 2S on the supporting
material. All samples showed slight increases in the IC since ultrasonic
treatment acts in the separation of the nanoparticles. Longer times may
have resulted in an entanglement of the fibers, which resulted in dis-
turbance of the crystalline region. The FT-IR spectra analysis also al-
lows a pre-selection of the samples based on the crystallinity: 2 h of
milling and 20 min of ultrasonication.
Similar to observed in the FTIR analysis, Raman frequency, in-
tensity, and band shape of the vibrations may vary between the nano-
celluloses due to the differences in the chemical structure, interaction
between the biopolymer chains, and its organization. In most cases, the
peak sharpness and intensities are associated with the crystallinity of
NC in which, increase in peak intensities increases in the crystallinity of
NC as shown in Fig. 4a–b. Crystallinity percentage after ultrasonic ir-
radiations drastically increase in all the study samples, as presented in
Table 2; this may be due to the disruption of amorphous domains to
yield nanocellulose because of ultrasonic irradiation.
Raman spectra of cellulose from eight different treatments were
collected and compared to the literature data. Cellulose is a poly-
saccharide thus has characteristics Raman bands from 300 to 3000 cm-
1
. The main cellulosic peaks are indicated with arrows, and corre-
sponding peaks are ˜380 and ˜460 cm-1: CeCeC, C–O, CeC–O ring
deformation. The band at 900 cm-1 is associated with the C–H ring of
cellulose and is the disorder of the cellulosic structure (amorphous re-
gion) [33]. Similarly, 1340 cm-1: −CH2, HeCeC, HeOeCe, and
C–OeH; 1456 cm-1: HeC and HeOeC bending. Raman bands from
1300 to 1500 cm-1 corresponding to −CH2 and −CH2OH deformations
are also reported [33]. The Raman bands at 2889 cm-1: CH and CH2
corresponding to stretching and bands at 3286−3402 cm-1: OeH
stretch [42].
Raman technique can result in estimate values of crystallinity, as
presented in Fig. 4a–b. This value can be influenced by hemicelluloses
(as impurities), and therefore other measurement methods need to be
conducted and compared. of crystallinity by the ratio of cellulose's
Raman bands at 380 and 1096 cm-1. Also, the region of the sample or
the preparation of the sample can influence the obtained values, and
this is the reason why the results of crystallinity measurements by
different researches have been, at times, contradictory [34].
It was possible to sequence the crystallinity for the milled samples as
2 M ˜ 3 M > 1 M > 4 M; the results indicate that the best condition of
milling time considering only the crystallinity of the samples is two
hours. Evaluating the ultrasonicated samples, all of them showed sig-
nificant increases in the crystallinity values. The defibrillation effect
can justify the increase in the IC values after the high-intensity ultra-
sound: when the ultrasonic waves arrive at the sample, they penetrate
the crystalline structure of the cellulose and result in the separation of
R.R. Ferreira, et al. Materials Today Communications 22 (2020) 100755

Fig. 2. Infrared spectra of (a) Eucalyptus (ES) and 1 M, 2 M, 3 M and 4 M samples and zoom in the region 1800 to 1400 cm-1 and 1390 to 600 cm-1; and (b) full spectra
for the ultrasound-20 min samples.

Table 2 this technique is the hydrodynamic radius (RH) of a sphere having the
Crystallinity values obtained from FTIR and Raman spectra. same diffusion coefficient as the rod-like cellulose nanoparticles
Samples Apparent IC (%) (FTIR) 1369/1315 Crystallinity (%) (Raman)
[8,43,44]. Fig. 3S shows the curves of the size distribution and in-
formation about the maximum, minimum, and average sizes of the
1M 98.4 49.8 milled samples.
1 M-20U 99.3 57.9 Comparing the milled samples, in Fig. 3S, it was possible to observe
2M 99.4 59.9
2 M-20U 99.7 77.0
that the RH Average was the lowest for 2 h of ball mill, followed for 3 h,
3M 88.1 60.3 4 h, and 1 h. The milling time of the 1 h was not enough to break the
3 M-20U 89.9 73.5 fibers into nanostructures; the higher times (3 and 4 h) show a trend of
4M 98.1 17.7 increase de dimensions. The best milling time was 2 h. According to
4 M-20U 99.9 66.3
Sofla et al., longer grinding times may result in the growth of a fibrous
layer of cellulose around the grinding balls. This fact decreases the
impact power of the balls on the particles and results in larger sizes
the nanofibrils. This separation, associated with the high shear force
[45,46]. This behavior justifies the larger particle sizes of the samples
present in the solution, breaks the amorphous regions and isolates the
3 M and 4 M.
crystalline ones, resulting in more separated and crystalline NCs. The
In all DLS measures, it was verified that the sample with 20 min of
more crystalline the sample ground, the easier the separation of na-
ultrasound showed the lowest RH averages. Most of the particles also
nofibrils, and therefore the best results among the ultrasonic samples
presented maximum sizes in the micrometric range is indicative of the
are those for 2 h of milling.
more elongated morphology of the particles (nanofibrils – Fig. 6)
[7,47]. This size trend can be observed in Fig. 5, which correlates the
3.2. Preliminary evaluation of sizes and shapes and morphological analysis average particle size and ultrasound time.
Correlations with the SLS technique and morphological analyses
The milling was conducted in four different times: 1, 2, 3, and 4 h, were needed to confirm the shape of the NCs. In this way, SLS analysis
and the effects of the different times were evaluated dimensionally. The was conducted, and Table 3 presents the obtained values of the hy-
DLS and SLS analysis allow a preliminary evaluation of the NC’ sizes drodynamic radius (RH), rotation radius (RG), distribution of sizes, and
and shapes. The DLS measures the translational diffusion coefficient of identification of the geometric shape of objects (RG/RH).
particles in water, which undergo Brownian motion. The value given by With the SLS data, it is possible to evaluate the parameter ρ, related

5
R.R. Ferreira, et al.
Fig. 3. Challenges of ball mill in the nanocellulose (NC) isolation, comparison between the real and the aimed scenario, and the main techniques that can be
employed to analyze each NC characteristic.
to the RG (obtained from SLS values) and RH (obtained from DLS va-
lues). The RG is a parameter defined as an average root mean squared
distance from the center of the mass of a macromolecule. This para-
meter allows a correlation between the RG/RH and the morphology of
the systems studied. It is a little-used technique for understanding cel-
lulosic systems because of its complexity; however, it provides basic
information in a short time.
According to Stetefeld et al., RG/RH ˜1 can be attributed to elon-
gated shapes [48]. It is possible to observe that there is much different
information that needs to be compilated and corroborated with other
morphological techniques. In the case of NCs, many shapes can occur,
such as spheres, fibrils, and rod-like structures [31,49–51]. So, there is a
limitation in the precise estimation of the shapes, resulting in approx-
imate values.
According to the values, the time of 20 and 30-min ultrasound are
considered adequate to obtain an elongated shape. The sample 1 M-30U
Fig. 4. Raman spectra of (c) milled and (d) ultrasonicated samples.
6
Materials Today Communications 22 (2020) 100755
(ρ = 1.48) and 2 M-20U (ρ = 1.22) are considered as an excellent
morphology due to its ρ, which fits into the fibrillar morphology be-
cause they present the ratio of ρ > 1 [48].
To corroborate and complement the DLS and SLS results, the mor-
phological analysis was conducted. Fig. 6a shows the FESEM images of
the milled and ultrasonicated samples. The FESEM images validate the
changes in the disintegration of the fibers, and the best samples were
evaluated by TEM images, as shown in Fig. 6b that is the 2 h milled
sample and 2 M assisted by 20 min of high-intensity ultrasound [70].
The best samples were determined according to the previous results
presented in Fig. 4S.
The use of FESEM is limitated by the equipment conditions, being
necessary to dry the samples, which results in agglomeration and
overestimation of sizes. NC usually aggregate into bundles during the
drying process, and that is why FESEM images are not so easy to see and
individual nanoparticles could not be observed. Also, the surface of the
R.R. Ferreira, et al.
1M
2M
400 3M
4M
300
Size (nm)
200
100
0 10 20 30
Ultrasound Time (min)
Fig. 5. Average hydrodynamic sizes of the NC vs. ultrasound time, measured by
DLS analysis.
sample was coated with gold to obtain the images [71], and the coating
broads the nanosized structure. However, this technique provided some
interesting observations about the samples.
The use of the ball mill resulted in a non-uniform and random
breaking of the cellulose fibers. Some fibers have better defibrillation,
while others remain micrometric in size. In some FE-SEM images, it is
possible to observe some regions have a quasi-circular shape, for ex-
ample, in 4 M, 1 M-20U, 2 M-20U, and 4 M-20U. It is associated with the
ball milling effect due to the breakage of the fibers along the direction
perpendicular to the fiber axis [60].
Based on all other results observed during this work, it was con-
cluded that the 2 M sample is the one with the most potential to obtain
nanocellulose, and the use of 20 min of ultrasound was the time that
presented the best results. In this way, the morphology verification
using TEM was performed for these samples (Fig. 6b). It is interesting to
observe that in sample 2 M, which was in solution and, therefore, in a
dispersed state when compared to FESEM, presented diameters of ap-
proximately 400−600 nm and length of ˜5 μm, which result in an as-
pect ratio (L/D) of ˜8. After the ultrasound, these samples presented a
drastic reduction of diameters, with an average value of ˜25 nm and
length of ˜3 μm, which result in an aspect ratio (L/D) of ˜120 [71]. By
these results, it is possible to conclude that the morphology of the ob-
tained NCs are nanofibers.
In Fig. 6c is presented a scheme that represents the proposed ex-
planation about the ultrasound effect on the NCs. The mechanical
grinding by ball mill resulted in a random and irregular break due to
the different forces that occur in the fibers: shear, friction, and com-
pression. By using ultrasound as a combined method, ultrasound waves
generate more defibrillation of fibers, which are already partially ac-
cessible, and these results are more visible with TEM images. It results
in greater uniformity of sizes, smaller diameters, and more stability of
the suspension.
3.3. Zeta potential
Zeta potential (δ) was measured to evaluate the electrostatic stabi-
lity of the NCs, and the results are presented in Table 3. Values less
than ± 15.0 mV represent the tendency to agglomeration and greater
than ± 30.0 mV means that the samples show electrostatic charges
[30,52,53]. The sample 1 M was the only that showed electrostatic
instability, i.e., a tendency of aggregation and flocculation due to the
low presence of free hydroxyl groups. It probably occurred due to the
larger particle sizes, which mean lower surface area and lower avail-
ability of free hydroxyls; also, the gravitational forces act and tend to
flocculate large particles. The other milled samples showed high δ
7
Materials Today Communications 22 (2020) 100755
values. Considering the δ values of the samples (Table 3), after the
ultrasound, the samples 4 M reduced the stability of the suspension, for
3 M's there is no significant change in this properties, and the 2 M
samples increased the stability for longer times of ultrasound, whereas
the 1 M’s the same behavior was observed. The sample 2 M-20U stood
out with ˜60 mV, which means that there is possibly a greater avail-
ability of free hydroxyls that result in repulsive electrostatic charges
[54,55]. The 2 h ball-milled, and 20-min ultrasound time was selected
as the ideal time in this study, considering dimensional and surface
characterization and the shortest time of use of ultrasound.
The suspensions were monitored as a function of time at 0, 10, and
30 min, as shown in Fig. 7. The milled samples show large particle sizes
and aggregation tendency, being visible to the naked eye. After 10 min,
the samples were deposited at the bottom of the container. The larger
sizes and the superficial charges resulted in a high tendency to floc-
culation [56].
Samples that have passed through the ultrasound process showed
better dispersibility. It is probably because the samples have smaller
particle sizes since it is expected that the high-intensity ultrasonication
promotes a good disintegration of the cellulose fibers. In this case, the
high superficial area probably has high free hydroxyl contents, which
contributes to the dispersibility of the suspension [57,58].
3.4. X-ray diffraction
The crystalline behavior of the milled samples before and after the
high-ultrasonication was investigated using X-ray diffraction. Fig. 8a–b
show the XRD patterns, and all the samples follow a similar trend ex-
hibiting the principal crystal planes (200) and (004) at approximately
22.2 and 34.6°, respectively [22]. The small and broad peak at 34.5°
represents the contribution of (040) plane. The index of crystallinity
(IC) indicated that all the samples have a higher IC than the ES (raw);
this could be justified with the removal of the non-cellulosic compo-
nents (hemicellulose and lignin) after ultrasonic treatment [41]. Com-
paring with the IC calculated by FTIR analysis, it is possible to observe a
similar trend, with the ultrasonicated samples but has a slight differ-
ence in the crystallinity values. This may occur because there may still
be other lignocellulosic components present, and when used the FTIR
technique, this may interfere significantly in the calculated values [32].
In Fig. 8a it is possible to observe that the samples presented IC
values in the following order; 1 M > 4 M > 2 M > 3 M. The 1 M sample
was subjected to less milling time resulting in fewer shocks and shear
force. This demonstrates that despite reducing nanometer-scale parti-
cles, ball milling results in loss of crystallinity over time. One hypoth-
esis for the 4 M sample to have been more crystalline than 2 M and 3 M
is that there was enough breakdown of the cellulose crystalline region
that was exposed and making this into the amorphous region.
The use of high-intensity ultrasound methodology presented results
similar to those of grinding, as present in Fig. 8b. However, obtained
NC with ultrasonically treated samples exhibited homogeneous dis-
persion and disintegration into individual particles due to the disrup-
tion of hydrogen bonds. Nevertheless, some authors mentioned the
limitations of this method, mainly due to the underestimation of the
contribution of the amorphous region, overestimating the IC [41]. One
justification for the similarity between the IC values is that no acid was
used for the dissolution of the amorphous region, only breakage. As the
NC solutions do not undergo any washing, as reported in the experi-
mental part, there is no separation between amorphous and crystalline
regions. For a better understanding of the difference between methods,
the morphological analysis was used. Similar results were reported by
Chen et al. [59,60].
Based on the previous results of crystallinity and nanosizes, these
parameters were evaluated to determine the most favorable condition
to obtain NC. The compilated results are presented in Fig. 4S, in the
Supplementary Material, and show that the 2 h of milling (sample 2 M)
was considered the ideal grinding time. Also, concerning the ultrasound
R.R. Ferreira, et al.
Fig. 6. (a) FESEM images of the milled and ultrasonicate samples and (b) TEM micrographs of the best conditions: 2 M and 2 M-20U; (c) Scheme shows the main
consideration about the effect of the high-intensity ultrasonic.
times evaluated, the use of 20 min presented more uniform sizes and
smaller size distribution. Thus, the most effective combination of
methods was for 2 h of milling assisted with 20 min of ultrasound.
8
Materials Today Communications 22 (2020) 100755
3.5. Evaluation of the composition and eventual contamination of the NCs
The high-resolution XPS spectra were collected for the milled
samples and the ultrasonicated samples for 20 min (Fig. 8c–d), re-
spectively. It was observed that increasing the milling time there is the
presence of Al and Si. These elements are impurities associated with the
R.R. Ferreira, et al. Materials Today Communications 22 (2020) 100755

Table 3 milled samples. This difference in NC values is due to the chemical pre-
Structural parameters using light scattering techniques (RH values were ob- treatment and the combinatorial isolation method used to obtain NCs
tained from DLS and RG values from SLS), and zeta potential values. from the ES fibers, in which the ball mill allowed the partial removal of
Samples RH (nm) RG (nm) ρ (RG/RH) Zeta potential (mV) oxygen from the NCs while the high-intensity ultrasound caused a
closer increase of pure NC (0.83), as reported in the literature [62].
1M 2840.0 338.3 0.12 −24.7
1 M-10U 453.6 132.6 0.29 −51.7
1 M-20U 361.4 156.6 0.43 −40.8
3.6. Thermogravimetric analysis (TGA)
1 M-30U 113.4 167.3 1.48 −40.9
2M 246.8 221.4 0.90 −43.5
2 M-10U 450.9 173.5 0.39 −42.1 The thermal analysis gives information about the thermal behavior
2 M-20U 145.9 177.3 1.22 −59.3 of the ES and NC derived from it as well as possible high-temperatures
2 M-30U 826.0 150.8 0.18 −53.5 applications of NC [22]. The results obtained by the thermogravimetric
3M 2600.0 306.8 0.19 −45.0
3 M-10U 544.6 197.5 0.36 −45.7
analysis were plotted as thermogravimetric and its derivative curves, as
3 M-20U 488.8 156.1 0.32 −39.3 in Fig. 9a–b.
3 M-30U 277.6 167.7 0.60 −45.7 The TGA curves maintained a similar pattern for all the samples,
4M 240.1 207.3 0.86 −45.7 which is in agreement with the conventional curves of NC [53,63–68].
4 M-10U 301.4 131.1 0.44 −48.1
All these samples showed two main events of percentage weight
4 M-20U 374.5 143.4 0.38 −40.7
4 M-30U 216.8 154.3 0.71 −40.5 loss. The first is related to the breakage of inter- and intramolecular
hydrogen bonds corresponding to water evaporation [30,58]. The
second event of weight loss is related to the thermal degradation of
jar and balls composition (SiO2 and alumina), and higher times of cellulose. The ES sample had slightly higher thermal degradation
milling resulted in higher contents of impurities. The O/C ratio was temperatures than the NCs. Nanoparticles have larger surface areas
calculated, and the obtained values are 0.39, 0.49, 0.54, 0.68, and 0.75 than ES, which may have catalyzed thermal degradation. All the milled
for the samples ES, 1 M, 2 M, 3 M, and 4 M, respectively. The theoretical samples showed very similar Tmax, even at high mechanical milling
value of pure cellulose is 0.83 [61]. times, which indicates that all the structures are identical, and similar
Alumina is the main composition of the jar and spheres; Silicon energies are required to thermal degradation of the cellulose.
dioxide is the "contaminant" mostly found in the balls. Interestingly, as There was an increase in decomposition temperature by 10 °C in all
milling time increased, the aluminum concentration was constant and the ultrasonically treated samples than ball-milled samples. This small
low; however, there was an increase in silicon content due to the wear change could be due to the more ordered and individualized structures
of the balls, which mainly occurs in the region of impurities, resulting which corroborate with the previously presented crystallinity results.
in contamination of the cellulose. The use of other materials as grinding Another possibility is that the surface area of the NC is very high, which
agents such as high alumina or zirconic spheres could significantly re- means that the heat can be disseminated very quick, and it can decrease
duce contamination but would result in an increase of approximately 20 the thermal decomposition temperature of the samples.
% in cost. This behavior is another indication of the efficiency of the combined
For the ultrasonicated samples, the values are 0.61, 0.64, 0.61, and methods since the high temperatures are justified by the higher energy
0.79 for the samples 1 M-20U, 2 M-20U, 3 M-20U, and 4 M-20U, re- necessary for the thermal degradation of the nanocelluloses [69]. These
spectively. The ultrasound samples showed higher O/C values than the results are promising and demonstrate that the combination of the two
insulation methods may favor applications that require higher

Fig. 7. Samples suspensions at the times of 0, 10, and 30 min.

9
R.R. Ferreira, et al.
Fig. 8. X-ray diffraction patterns of (a) milled samples and (b) ultrasonicated samples; XPS survey spectra (0–1400 eV) for (c) eucalyptus sawdust and milled NCs and
(d) ultrasonicated samples, showing the representative C 1s (286 eV) and O 1s (532 eV) photoemission peaks.
processing temperatures.
3.7. Final remarks
Nanocellulose, with exceptional intrinsic properties such as ex-
cellent thermal and dimensional stability, crystallinity, and stiffness,
has been prominent in the world bioeconomy as a potential substitute
for carbon and graphene nanotubes. This is because this material has
high added value and is a class of compatible and renewable biona-
nomaterials. This reflects a worldwide race for NC production with high
productivity, process efficiency, and low environmental impacts.
Considering industrial production or pilot-scale, some companies may
be cited. VTT Finland is developing a project to obtain NC with enzy-
matic processes; Paperlogic has a production capacity of 2 tons per day;
Fibria Insights is already investing in innovative projects to develop
viable processes for obtaining NCs; In addition, numerous start-ups are
springing up around the world to develop these high-quality, high-
quality advanced materials [29,72–74].
It should be noted, however, that many companies, in addition to
targeting high-capacity production processes, have also shown a keen
interest in developing more environmentally friendly techniques [75].
Although numerous researches are being conducted, there is still a
limitation on mechanical methods due to commonly used equipment.
Thus, the combinatorial approach presented in this work stands out for
using common equipment in other industrial areas (such as paints and
ceramic suspensions), without using toxic reagents. There is a clear and
growing trend in the development of methods and processes aimed at
lowering prices and increasing nanocellulose productivity, which has
proven to be a "material of the future".
10
Materials Today Communications 22 (2020) 100755
4. Conclusions
Nanocellulose (NC) was successfully isolated from eucalyptus saw-
dust using a combination of ball mill and high-intensity ultrasonication
methods; all the proposed milling times (1, 2, 3, and 4 h) resulted in
cellulosic particles in the nanometric scale. The use of a combinatorial
method to overcome the challenges encountered in mechanical
grinding without the use of chemical reagents was successful; the ul-
trasound
allows defibrillation of the nanofibres by reducing NC diameters, as
well as providing high power output, concentrated energy transfer, low
cost and ecological character. The 2 h-milling times presented smaller
size, following by 3 M < 4 M < 1 M. After the milling, the ultrasound
irradiation was conducted in 10, 20, and 30 min. The 20 min samples
showed the smaller sizes and high electrostatic stability when compared
to the other evaluated times; and according to static light scattering,
this sample has random morphology with elongated form, which is
desired for NC.
The crystallinity values (CI), determined by FTIR, Raman, and XRD,
show a similar trend. The 2 h-milled sample presented a high crystal-
linity, and the CI values followed the order 2 h˜3 h > 4 h > 1 h; longer
times (4 h) may have a disturbance in the crystalline region; low times
(1 h) was not enough to break and remove the amorphous region. Also,
the use of ultrasound increased the thermal stability, which can be
justified by the increase in the crystallinity of the samples.
It was verified that the mechanical grinding results in a random and
irregular break of the cellulose fibers due to the shear and friction
forces. So, in the combined method, the ultrasound waves generate
more defibrillation of nanofibers, decreasing the NC’ diameters. Short
ultrasonic times are not enough for the separation of the nanofibers,
whereas excessive times induce a disorder and entanglement of the NC.
Also, they produce fibers with fewer contaminants, according to XPS
R.R. Ferreira, et al.
Fig. 9. Thermal behavior of the ES sample and their prepared nanoparticles. ES and their nanocelluloses thermograms without (a) and with (b) the use of ultrasound.
The inserts are the corresponding DTG curves.
results.
This study allowed an in-depth understanding of the combining
method of grinding and ultrasound, which is a completely toxic-free
and widely applicable method. The presented results suggest that the
proposed method has a high potential and alternative way to unravel
highly dispersive NC from natural resources. Also, the obtained nano-
particles are expected to show high potential as nanocellulosic re-
inforcement into a polymeric matrix, providing value-added materials
with superior performance and biodegradable character. The applica-
tions are innumerous, such as agricultural plastics, packaging, oil
emulsification, and others.
Declaration of Competing Interest
The authors declare that there are no conflicts of interest.
Acknowledgments
The authors thank the financial support provided by FAPESP (2018/
11277-7 and 2018/25239-0), CNPq (305819/2017-8), CAPES (Code
001), NSF-CREST #1735971, and the Multiuser Experimental Center of
the Federal University of ABC (CEM-UFABC).
11
Materials Today Communications 22 (2020) 100755
Appendix A. Supplementary data
Supplementary material related to this article can be found, in the
online version, at doi: https://doi.org/10.1016/j.mtcomm.2019.
100755.
References
[1] C.L. Pirich, G.F. Picheth, J.P.E. Machado, C.N. Sakakibara, A.A. Martin, R.A. de
Freitas, M.R. Sierakowski, Influence of mechanical pretreatment to isolate cellulose
nanocrystals by sulfuric acid hydrolysis, Int. J. Biol. Macromol. 130 (2019)
622–626, https://doi.org/10.1016/j.ijbiomac.2019.02.166.
[2] F. and agriculture organization of the U.N. FAO, Global data on forest plantations
resources, (n.d.). http://www.fao.org/3/Y2316E/y2316e0b.htm.
[3] J.O. Brito, F.G. Silva, M.M. Leão, G. Almeida, Chemical composition changes in
eucalyptus and pinus woods submitted to heat treatment, Bioresour. Technol. 99
(2008) 8545–8548, https://doi.org/10.1016/j.biortech.2008.03.069.
[4] P. Indu, Dados Estatísticos, (2013), pp. 1–3.
[5] P. Phanthong, G. Guan, Y. Ma, X. Hao, A. Abudula, Effect of ball milling on the
production of nanocellulose using mild acid hydrolysis method, J. Taiwan Inst.
Chem. Eng. 60 (2016) 617–622, https://doi.org/10.1016/j.jtice.2015.11.001.
[6] R. da Silva, M.A. Woehl, C.N. Sakakibara, A.A. Martin, M.R. Sierakowski,
G.F. Picheth, C.L. Pirich, C.F. de Souza, R.A. de Freitas, Bacterial cellulose in bio-
medical applications: a review, Int. J. Biol. Macromol. 104 (2017) 97–106, https://
doi.org/10.1016/j.ijbiomac.2017.05.171.
[7] T. Kos, A. Anžlovar, M. Kunaver, M. Huskić, E. Žagar, Fast preparation of nano-
crystalline cellulose by microwave-assisted hydrolysis, Cellulose 21 (2014)
2579–2585, https://doi.org/10.1007/s10570-014-0315-2.
[8] M. Kunaver, A. Anžlovar, E. Žagar, The fast and effective isolation of nanocellulose
from selected cellulosic feedstocks, Carbohydr. Polym. 148 (2016) 251–258,
R.R. Ferreira, et al.
https://doi.org/10.1016/j.carbpol.2016.04.076.
[9] S. Bee, A. Hamid, H.Vc. of L.B. to N.: S. and C.P.C. Lee, S.B. a Hamid, S.K. Zain, rsion
of Lignocellulosic Biomass to Nanocellulose :, 2014 (2015). doi:10.1155/2014/
631013.
[10] A.G. de Souza, D.B. Rocha, F.S. Kano, D.dosS. Rosa, Valorization of industrial paper
waste by isolating cellulose nanostructures with different pretreatment methods,
Resour. Conserv. Recycl. 143 (2019) 133–142, https://doi.org/10.1016/j.
resconrec.2018.12.031.
[11] W. Pan, X. Zhang, F. Niu, M. Li, J. Li, J. Yang, Q. Huang, The characteristic and
dispersion stability of nanocellulose produced by mixed acid hydrolysis and ultra-
sonic assistance, Carbohydr. Polym. 165 (2017) 197–204, https://doi.org/10.1016/
j.carbpol.2017.02.048.
[12] Z. Karim, S. Afrin, Q. Husain, R. Danish, Necessity of enzymatic hydrolysis for
production and functionalization of nanocelluloses, Crit. Rev. Biotechnol. 37 (2017)
355–370, https://doi.org/10.3109/07388551.2016.1163322.
[13] C.J. Garvey, P. Memmott, P.K. Annamalai, D.J. Martin, E. Jiang, N. Amiralian, High
aspect ratio nanocellulose from an extremophile spinifex grass by controlled acid
hydrolysis, Cellulose 24 (2017) 3753–3766, https://doi.org/10.1007/s10570-017-
1379-6.
[14] T. Theivasanthi, F.L. Anne Christma, A.J. Toyin, S.C.B. Gopinath, R. Ravichandran,
Synthesis and characterization of cotton fiber-based nanocellulose, Int. J. Biol.
Macromol. 109 (2018) 832–836, https://doi.org/10.1016/j.ijbiomac.2017.11.054.
[15] J. Hu, D. Tian, S. Renneckar, J.N. Saddler, Enzyme mediated nanofibrillation of
cellulose by the synergistic actions of an endoglucanase, lytic polysaccharide
monooxygenase (LPMO) and xylanase, Sci. Rep. 8 (2018) 1–15, https://doi.org/10.
1038/s41598-018-21016-6.
[16] N. Kruer-Zerhusen, B. Cantero-Tubilla, D.B. Wilson, Characterization of cellulose
crystallinity after enzymatic treatment using Fourier transform infrared spectro-
scopy (FTIR), Cellulose 25 (2018) 37–48, https://doi.org/10.1007/s10570-017-
1542-0.
[17] M. Nuruddin, M. Hosur, M. Jamal Uddin, D. Baah, S. Jeelani, A novel approach for
extracting cellulose nanofibers from lignocellulosic biomass by ball milling com-
bined with chemical treatment, J. Appl. Polym. Sci. 133 (2016), https://doi.org/10.
1002/app.42990.
[18] M.C. Guimara, Evaluation of the Effect of Processing Conditions on the Impact
Properties of, (2017), https://doi.org/10.1177/0021998316688338.
[19] S. Mondal, Preparation, properties and applications of nanocellulosic materials,
Carbohydr. Polym. 163 (2017) 301–316, https://doi.org/10.1016/j.carbpol.2016.
12.050.
[20] F.S. Kano, A.G. de Souza, D.dosS. Rosa, Variation of the milling conditions in the
obtaining of nanocellulose from the paper sludge, Matéria (Rio Janeiro). 24 (2019),
https://doi.org/10.1590/s1517-707620190003.0719.
[21] C.C. Piras, S. Fernández-Prieto, W.M. De Borggraeve, Ball milling: a green tech-
nology for the preparation and functionalisation of nanocellulose derivatives,
Nanoscale Adv. 1 (2019) 937–947, https://doi.org/10.1039/c8na00238j.
[22] C.M.E. Xuran, L. Min, W.Y. Huang, C.M. Ewulonu, Á.X. Liu, Á.M. Wu, Á.Y. Huang,
X. Liu, Ultrasound-assisted mild sulphuric acid ball milling preparation of lig-
nocellulose nanofibers (LCNFs) from sunflower stalks (SFS), Cellulose 4 (2019),
https://doi.org/10.1007/s10570-019-02382-4.
[23] Z. Ling, T. Wang, M. Makarem, M. Santiago Cintrón, H.N. Cheng, X. Kang,
M. Bacher, A. Potthast, T. Rosenau, H. King, C.D. Delhom, S. Nam, J. Vincent
Edwards, S.H. Kim, F. Xu, A.D. French, Effects of ball milling on the structure of
cotton cellulose, Cellulose 26 (2019) 305–328, https://doi.org/10.1007/s10570-
018-02230-x.
[24] Y. Zheng, Z. Fu, D. Li, M. Wu, Effects of ball milling processes on the microstructure
and rheological properties of microcrystalline cellulose as a sustainable polymer
additive, Materials (Basel) 11 (2018) 1–13, https://doi.org/10.3390/ma11071057.
[25] H.P.S. Abdul Khalil, Y. Davoudpour, M.N. Islam, A. Mustapha, K. Sudesh,
R. Dungani, M. Jawaid, Production and modification of nanofibrillated cellulose
using various mechanical processes: a review, Carbohydr. Polym. 99 (2014)
649–665, https://doi.org/10.1016/j.carbpol.2013.08.069.
[26] A. Barakat, C. Mayer-Laigle, A. Solhy, R.A.D. Arancon, H. De Vries, R. Luque,
Mechanical pretreatments of lignocellulosic biomass: towards facile and en-
vironmentally sound technologies for biofuels production, RSC Adv. 4 (2014)
48109–48127, https://doi.org/10.1039/c4ra07568d.
[27] R.J. Moon, A. Martini, J. Nairn, J. Simonsen, J. Youngblood, Cellulose
Nanomaterials Review: Structure, Properties and Nanocomposites, (2011), https://
doi.org/10.1039/c0cs00108b.
[28] J. Desmaisons, E. Boutonnet, M. Rueff, A. Dufresne, J. Bras, A new quality index for
benchmarking of different cellulose nanofibrils, Carbohydr. Polym. 174 (2017)
318–329, https://doi.org/10.1016/j.carbpol.2017.06.032.
[29] M.S. Reid, M. Villalobos, E.D. Cranston, Benchmarking cellulose nanocrystals: from
the laboratory to industrial production, Langmuir 33 (2017) 1583–1598, https://
doi.org/10.1021/acs.langmuir.6b03765.
[30] G.F. de Lima, A.G. de Souza, D.S. Rosa, Effect of adsorption of polyethylene glycol
(PEG), in aqueous media, to improve cellulose nanostructures stability, J. Mol. Liq.
268 (2018) 415–424, https://doi.org/10.1016/j.molliq.2018.07.080.
[31] A.G. De Souza, F.S. Kano, J.J. Bonvent, D.S. Rosa, Cellulose Nanostructures
Obtained From Waste Paper Industry : A Comparison of Acid and Mechanical
Isolation Methods, (2017), pp. 1–6.
[32] C. Lee, K. Dazen, K. Kafle, A. Moore, D.K. Kohnson, S. Park, S.H. Kim, Correlations
of Apparent Cellulose Crystallinity Determined by XRD, NMR, IR, Raman, and SFG
Methods, in: Adv. Polym. Sci. Springer International Publishing, Switzerland, 2015,
pp. 1–34, https://doi.org/10.1007/12.
[33] U.P. Agarwal, Raman spectroscopy in the analysis of cellulose nanomaterials,
Nanocelluloses Their Prep. Prop. Appl. ACS Sympos (2017) 75–90, https://doi.org/
12
Materials Today Communications 22 (2020) 100755
10.1021/bk-2017-1251.ch004.
[34] U.P. Agarwal, Raman Spectroscopy of CNC- and CNF-based nanocomposites,
Handb. Nanocellulose Cellul. Nanocompos. (2017) 609–625, https://doi.org/10.
1002/9783527689972.ch18.
[35] A. Jäger, E. Jäger, F. Surman, A. Höcherl, B. Angelov, K. Ulbrich, M. Drechsler,
V.M. Garamus, C. Rodriguez-Emmenegger, F. Nallet, P. Štěpánek, Nanoparticles of
the poly([N-(2-hydroxypropyl)]methacrylamide)-b-poly[2-(diisopropylamino)ethyl
methacrylate] diblock copolymer for pH-triggered release of paclitaxel, Polym.
Chem. 6 (2015) 4946–4954, https://doi.org/10.1039/c5py00567a.
[36] H.L. Yi, L.J. Lai, J.S. Jiang, C.C. Hua, Ethylcellulose colloids incubated in dilute
solution, J. Phys. Chem. B 121 (2017) 638–648, https://doi.org/10.1021/acs.jpcb.
6b09976.
[37] S. Nam, A.D. French, B.D. Condon, M. Concha, Segal crystallinity index revisited by
the simulation of X-ray diffraction patterns of cotton cellulose Iβ and cellulose II,
Carbohydr. Polym. 135 (2016) 1–9, https://doi.org/10.1016/j.carbpol.2015.08.
035.
[38] L. Segal, J.J. Creely, A.E. Martin, C.M. Conrad, An empirical method for estimating
the degree of crystallinity of native cellulose using the X-ray diffractometer, Text.
Res. J. 29 (1959) 786–794, https://doi.org/10.1177/004051755902901003.
[39] R.M. de Sá, C.S. de Miranda, N.M. José, Preparation and characterization of na-
nowhiskers cellulose from fiber arrowroot (Maranta arundinacea), Mater. Res. 18
(2015) 225–229, https://doi.org/10.1590/1516-1439.366214.
[40] R.A. Ilyas, S.M. Sapuan, M.R. Ishak, Isolation and characterization of nanocrystal-
line cellulose from sugar palm fibres (Arenga Pinnata), Carbohydr. Polym. (2017),
https://doi.org/10.1016/j.carbpol.2017.11.045.
[41] F. Kallel, F. Bettaieb, R. Khiari, A. García, J. Bras, S.E. Chaabouni, Isolation and
structural characterization of cellulose nanocrystals extracted from garlic straw
residues, Ind. Crops Prod. 87 (2016) 287–296, https://doi.org/10.1016/j.indcrop.
2016.04.060.
[42] J.S. Lupoi, E. Gjersing, M.F. Davis, Evaluating lignocellulosic biomass, its deriva-
tives, and downstream products with Raman spectroscopy, Front. Bioeng.
Biotechnol. 3 (2015) 1–18, https://doi.org/10.3389/fbioe.2015.00050.
[43] Y. Boluk, C. Danumah, Analysis of cellulose nanocrystal rod lengths by dynamic
light scattering and electron microscopy, J. Nanopart. Res. 16 (2014), https://doi.
org/10.1007/s11051-013-2174-4.
[44] Y. Mao, K. Liu, C. Zhan, L. Geng, B. Chu, B.S. Hsiao, Characterization of nano-
cellulose using small-angle neutron, X-ray, and dynamic light scattering techniques,
J. Phys. Chem. B 121 (2017) 1340–1351, https://doi.org/10.1021/acs.jpcb.
6b11425.
[45] V.K. Baheti, R. Abbasi, J. Militky, Ball milling of jute fibre wastes to prepare na-
nocellulose, World J. Eng. 9 (2012) 45–50.
[46] M.R.K. Sofla, R.J. Brown, T. Tsuzuki, T.J. Rainey, A comparison of cellulose na-
nocrystals and cellulose nanofibres extracted from bagasse using acid and ball
milling methods, Adv. Nat. Sci. Nanosci. Nanotechnol. 7 (2016), https://doi.org/
10.1088/2043-6262/7/3/035004.
[47] A.N. Frone, D.M. Panaitescu, I. Chiulan, C.A. Nicolae, Z. Vuluga, C. Vitelaru,
C.M. Damian, The effect of cellulose nanofibers on the crystallinity and nanos-
tructure of poly(lactic acid) composites, J. Mater. Sci. 51 (2016) 9771–9791,
https://doi.org/10.1007/s10853-016-0212-1.
[48] J. Stetefeld, S.A. McKenna, T.R. Patel, Dynamic light scattering: a practical guide
and applications in biomedical sciences, Biophys. Rev. 8 (2016) 409–427, https://
doi.org/10.1007/s12551-016-0218-6.
[49] T.T.T. Ho, T. Zimmermann, R. Hauert, W. Caseri, Preparation and characterization
of cationic nanofibrillated cellulose from etherification and high-shear disintegra-
tion processes, Cellulose 18 (2011) 1391–1406, https://doi.org/10.1007/s10570-
011-9591-2.
[50] L. Zhang, T. Tsuzuki, X. Wang, Preparation of cellulose nanofiber from softwood
pulp by ball milling, Cellulose 22 (2015) 1729–1741, https://doi.org/10.1007/
s10570-015-0582-6.
[51] R. Prathapan, R. Thapa, G. Garnier, R.F. Tabor, Modulating the zeta potential of
cellulose nanocrystals using salts and surfactants, Colloids Surf. A Physicochem.
Eng. Asp. 509 (2016) 11–18, https://doi.org/10.1016/j.colsurfa.2016.08.075.
[52] Y. Xu, A.D. Atrens, J.R. Stokes, Rheology and microstructure of aqueous suspen-
sions of nanocrystalline cellulose rods, J. Colloid Interface Sci. 496 (2017) 130–140,
https://doi.org/10.1016/j.jcis.2017.02.020.
[53] S. Naduparambath, J. T.V, V. Shaniba, S. M.P, A.K. Balan, E. Purushothaman,
Isolation and characterisation of cellulose nanocrystals from sago seed shells,
Carbohydr. Polym. 180 (2018) 13–20, https://doi.org/10.1016/j.carbpol.2017.09.
088.
[54] A. Pappu, M. Saxena, V.K. Thakur, A. Sharma, Facile extraction, processing and
characterization of biorenewable sisal fibers for multifunctional applications, J.
Macromol. Sci. Part A 1325 (2016), https://doi.org/10.1080/10601325.2016.
1176443.
[55] A.C. Corrêa, E.M. Teixeira, L.A. Pessan, L.H.C. Mattoso, Cellulose nanofibers from
curaua fibers, Cellulose 17 (2010) 1183–1192, https://doi.org/10.1007/s10570-
010-9453-3.
[56] C. Brinatti, J. Huang, R.M. Berry, K.C. Tam, W. Loh, Structural and energetic studies
on the interaction of cationic surfactants and cellulose nanocrystals, Langmuir 32
(2016) 689–698, https://doi.org/10.1021/acs.langmuir.5b03893.
[57] A. Kaboorani, B. Riedl, Surface modification of cellulose nanocrystals (CNC) by a
cationic surfactant, Ind. Crops Prod. 65 (2015) 45–55, https://doi.org/10.1016/j.
indcrop.2014.11.027.
[58] H. Lu, Y. Gui, L. Zheng, X. Liu, Morphological, crystalline, thermal and physico-
chemical properties of cellulose nanocrystals obtained from sweet potato residue,
Food Res. Int. J. 50 (2013) 121–128, https://doi.org/10.1016/j.foodres.2012.10.
013.
R.R. Ferreira, et al.
[59] W. Chen, H. Yu, Y. Liu, P. Chen, M. Zhang, Y. Hai, Individualization of cellulose
nanofibers from wood using high-intensity ultrasonication combined with chemical
pretreatments, Carbohydr. Polym. 83 (2011) 1804–1811, https://doi.org/10.1016/
j.carbpol.2010.10.040.
[60] R. Avolio, I. Bonadies, D. Capitani, M.E. Errico, G. Gentile, M. Avella, A multi-
technique approach to assess the effect of ball milling on cellulose, Carbohydr.
Polym. 87 (2012) 265–273, https://doi.org/10.1016/j.carbpol.2011.07.047.
[61] M.E. Achaby, Z. Kassab, A. Barakat, A. Aboulkas, Alfa fibers as viable sustainable
source for cellulose nanocrystals extraction: application for improving the tensile
properties of biopolymer nanocomposite films, Ind. CropsProd. 112 (2018)
499–510, https://doi.org/10.1016/j.indcrop.2017.12.049.
[62] S. Barazzouk, C. Daneault, Amino acid and peptide immobilization on oxidized
nanocellulose: spectroscopic characterization, Nanomaterials 2 (2012) 187–205,
https://doi.org/10.3390/nano2020187.
[63] E. Espinosa, J. Domínguez-Robles, R. Sánchez, Q. Tarrés, A. Rodríguez, The effect of
pre-treatment on the production of lignocellulosic nanofibers and their application
as a reinforcing agent in paper, Cellulose 24 (2017) 2605–2618, https://doi.org/10.
1007/s10570-017-1281-2.
[64] C. Xu, S. Zhu, C. Xing, D. Li, N. Zhu, H. Zhou, Isolation and properties of cellulose
nanofibrils from coconut palm petioles by different mechanical process, PLoS One
10 (2015) 1–11, https://doi.org/10.1371/journal.pone.0122123.
[65] N. Rambabu, S. Panthapulakkal, M. Sain, A.K. Dalai, Production of nanocellulose
fibers from pinecone biomass: Evaluation and optimization of chemical and me-
chanical treatment conditions on mechanical properties of nanocellulose films, Ind.
Crops Prod. 83 (2016) 746–754, https://doi.org/10.1016/j.indcrop.2015.11.083.
[66] Y.M. Zhou, S.Y. Fu, L.M. Zheng, H.Y. Zhan, Effect of nanocellulose isolation tech-
niques on the formation of reinforced poly(vinyl alcohol) nanocomposite films,
Express Polym. Lett. 6 (2012) 794–804, https://doi.org/10.3144/expresspolymlett.
2012.85.
[67] M.F. Rosa, E.S. Medeiros, J.A. Malmonge, K.S. Gregorski, D.F. Wood,
L.H.C. Mattoso, G. Glenn, W.J. Orts, S.H. Imam, Cellulose nanowhiskers from
13
Materials Today Communications 22 (2020) 100755
coconut husk fibers: Effect of preparation conditions on their thermal and mor-
phological behavior, Carbohydr. Polym. 81 (2010) 83–92, https://doi.org/10.
1016/j.carbpol.2010.01.059.
[68] S. Shankar, J.W. Rhim, Preparation of nanocellulose from micro-crystalline cellu-
lose: The effect on the performance and properties of agar-based composite films,
Carbohydr. Polym. 135 (2016) 18–26, https://doi.org/10.1016/j.carbpol.2015.08.
082.
[69] T.F. Meyabadi, F. Dadashian, G. Mir Mohamad Sadeghi, H. Ebrahimi Zanjani Asl,
Spherical cellulose nanoparticles preparation from waste cotton using a green
method, Powder Technol. 261 (2014) 232–240, https://doi.org/10.1016/j.powtec.
2014.04.039.
[70] J.C. Chandra, N. George, S.K. Narayanankutty, Isolation and characterization of
cellulose nanofibrils from arecanut husk fibre, Carbohydr. Polym. 142 (2016)
158–166, https://doi.org/10.1016/j.carbpol.2016.01.015.
[71] C. Liu, B. Li, H. Du, D. Lv, Y. Zhang, G. Yu, X. Mu, H. Peng, Properties of nano-
cellulose isolated from corncob residue using sulfuric acid, formic acid, oxidative
and mechanical methods, Carbohydr. Polym. 151 (2016) 716–724, https://doi.org/
10.1016/j.carbpol.2016.06.025.
[72] G. Chauve, J. Bras, Industrial Point of View of Nanocellulose Materials and Their
Possible Applications, Https://Umaine.Edu/Pdc/Cellulose-Nanomaterials/Order-
Nanocellulose/, (2014), pp. 233–252, https://doi.org/10.1142/
97898145664690014.
[73] T.P.D.C. of the U. of Maine, Order nanocellulose, Https://Umaine.Edu/Pdc/
Cellulose-Nanomaterials/Order-Nanocellulose/, (2018), pp. 4–5 https://umaine.
edu/pdc/cellulose-nano-crystals/.
[74] J. Miller, Nanocellulose State of The Industry, (2015).
[75] F. Piccinno, R. Hischier, S. Seeger, C. Som, Predicting the environmental impact of a
future nanocellulose production at industrial scale: application of the life cycle
assessment scale-up framework, J. Clean. Prod. 174 (2018) 283–295, https://doi.
org/10.1016/j.jclepro.2017.10.226.